At the beginning of this year I did an interview with Dahlia Mahlki of VMware research on the evolution of distributed ledger technologies (DLT) and some of the advances that VMware was contributing to in the industry. For a recap of that interview, click here.
There was a lot of buzz generated from that interview, and since then, I have been digging deeper into what is being worked on and what is the current state of any offering that may come from it. In April, I had the pleasure of speaking in detail to Michael DiPetrillo, Senior Director of Blockchain Technologies at VMware. This interview was recorded and the transcript is presented below.
In short, there are a lot of really transformational things going on in the industry and VMware is poised to be a management platform to enable companies to succeed in breaking new ground with blockchain based solutions, securely and quickly.
[DB]Welcome back to the technology emergency Podcast Episode 3. Enterprise blockchain on VMware. This is the second episode that we’ve done on VMware and blockchain. The first episode I did an interview with Dahlía Mahlki, founding member of the VMware Research. In this episode we’ll be speaking with senior director Mike DiPetrillo and we’ll be discussing some of the things that have happened since that last interview and the overall direction of blockchain technologies with VMWare. Hello Michael and welcome to the podcast. Thank you for taking time today to talk with us. Can you tell us a bit about yourself and your role at VMware?
[MD]Wonderful and thank you for having me on the podcast today. Appreciate that. So my name is Mike DiPretrillo, I’m a senior director at VMware. I concentrate on all things blockchain related at VMware today. So, in that role, actually I wear two different hats. One hat is on the research side of the world. This started as a research project at VMware and where we’ve been doing research in the blockchain space for the past four years. We are the world’s leading experts in kind of consensus protocols and distributed systems working at VMware, I’m lucky enough to have the opportunity to work with. And the other hat that I wear is on the product side. How do we take what we’ve been learning and doing in the research space and really bring it to the enterprises of the world to enable blockchain for enterprise use cases.
Two different hats at VMware. I’ve been at VMware for like 16 years always doing emerging technologies and one thing or another at here so blockchain is the next emerging technology that I’m working on and get the opportunity to play with.
[DB] Great. Thank you for that introduction. Now there are many applications of blockchain technology as both an enabler of solutions and as a solution itself. Is the VMware strategy to act as an enabler for enterprises to quickly evaluate and go to market with new business models or is that technology being integrated into the core product portfolio?
[MD] Absolutely. So it’s a little bit of both. We took the stance originally of looking at blockchain and how we make this more enterprise friendly. How do we make it so costs less energy both and you know, CPU hash power for example in energy costs. How do we make it more secure. How do we make it more scalable than really starting to look at focusing on day two operations. Once you have this blockchain up and running how do you really manage this thing from a day to day operations standpoint. How did you get it past the risk management department instead of your environment. These are all struggles that we see customers face over and over and over again that are mature in their blockchain adoption. So our main focus is really taking blockchain and making it useful for an enterprise in every different industry vertical that’s out there to build new applications on. As new applications maybe saving them money and they mean making them go faster in a particular market or a combination thereof.
But it’s really one of those two things that they’re concentrated on and then down the road we’re also taking it in integrated into our other products where it makes sense. Like we take NSX and wrap it around it for network security. And we look at managing it and operating it with wavefront we look at putting app defense on top of that from a security perspective. So there’s a lot of technologies not only from the VMWare stack but VMware is a member of Dell technologies as a whole. And so we have a partner companies that are out there so you get people like Dell Boomi saying how do we get information onto and off of the blockchain and tied into the business logic that’s out there. So it really is across Dell technologies that we look at this. But my job and what I focus on mostly on a day to day basis is at VMware, although I’m part of a group that shoves out across all Dell technologies. So really it’s an enabler for new business applications. While tied into our other VMware products.
[DB]OK. Now I understand that VMware has been contributing to the HyperLedger project. Specifically HyperLedger Fabric. What other HyperLedger projects have you also been contributing to?
[MD] Absolutely. So we give tribute to HyperLedger fabric. We partner with and work with some of the other projects that are in there like Quilt. Intel is a big partner of ours so we look at Sawtooth for example and in partnering with them a little bit on Sawtooth, but we don’t really contribute much there. Were actually the number one contributor in the Cello project which is really about managing and operating and deploying HyperLedger and that’s really kind of where our initial interest from one of our research teams in China came about.
But if you look at the VMware solution and what we’re what we’re building and what we’re deploying out there right now. We’ve heard from customers over and over again that there’s two things that really concerned them when they look at blockchain solutions. One of them is they don’t want to be tied into a particular hosting solution. So a lot of people are consuming blockchain from a hoster as a SaaS solution or as a hosted solution. They don’t want to be tied into a particular vendor there. And the second thing is they don’t want to be tied into a particular blockchain stack itself whether that be HyperLedger or Quorum or Corda or you know the Microsoft Azure Stack or any of these other stacks that are out there they really don’t want to be tied into one particular thing. So VMware has build a cross platform management security solution. So we have our own blockchain stack that we’ve developed inside of VMware. Our own consensus layer our own ledger layer, our own smart contracts layer. So that’s kind of the VMware Enterprise blockchain. Our management stack manages that. But it also manages HyperLedger inside that same interface, with that same look and feel right next to the VMware stack. And soon to be picking up a quorum on kind of the Ethereum side of the world as well. So we are multi-blockchain from a management to security and operations perspective already. And that’s one of the reasons we contribute to Cello. One of the reasons we contribute to fabric and other HyperLedger projects out there.
[DB]Now I’m curious. For VMware’s internal blockchain, have you moved away from proof of work or proof of stake to other distributed consensus mechanisms?
[MD] Yeah. VMware does something called Byzantine Fault Tolerance. So, if you look at kind of the history of the evolution of consensus protocols you know if you look like the Nakamoto consensus protocol which is really proof of work based system it really is very friendly for enterprises. A lot of CPU hashing power a lot of energy costs associated with that. And enterprises really don’t want to stand up a lot of CPU hashing power just to reach a consensus with each other. Right? You’ve got a couple of years pass and you get something called proof of stake which you highlighted initially created by peer to peer going but popularized by Ethereum. Most people think of Ethereum when they think of the Proof of Stake. A better system, you know faster but also takes capital real capital right. You have to invest it to get the gas to actually go and mine the blocks. So it is something that that enterprises typically don’t want to deal with is investing that capital. So if you look at Byzantine Fault Tolerance, you get to a true Byzantine Fault Tolerant or BFT engine you can start with as few as four nodes. It’s super lightweight. I mean I run our four nodes on my laptop with you know one CPU and like a gig of memory between the four nodes so it’s you know super lightweight that’s out there. You can scale this up. You know most environments don’t scale too well. You add more nodes you don’t necessarily get more nodes. So we create or get more throughput.
Our latest test was with 500 nodes spread across the world with 30000 transactions per second and finalized in 10 ms or less.
So we created something called Scalable Byzantine Fault Tolerance. Which is you know again high scale so simplified the protocol itself and how you can get to hundreds or thousands of nodes that are out there and you add more nodes you actually get more throughput. Our latest test was like with 500 nodes spread across the world with 30000 transactions per second and finalized in 10 milliseconds or less so high throughput high number of nodes low time to finality. Which is kind of the other thing that you look at in a consensus protocol. For how long until things are actually truly written. Now I don’t have to worry about forks in a Byzantine Fault Tolerant environment because it’s a voting algorithm. So that’s the path that we went down was going to the BFT route. We think it’s an evolutionary process inside of the environment and that’s you know where our path is leading.
[DB]Now is that something that can be applied across untrusted nodes or would that only be implemented in an isolated trusted node environment?
[MD]So we focus heavily on permissioned blockchains. Primarily because again we’re focused on those enterprise customers that are out there and they want to know which nodes are participating. Which are touching your data which nodes are voting on their data which are seeing their data. You know because most of those enterprise have to go through some sort of audit. One way or another at some point in time in a core part of any audit that comes out there is who is touching your data and seeing your data and where does your data live.
So if you can’t prove that or know that it’s become really hard past any sort of regulated audit this is you know one of the things that I think is is going to impact some of the public blockchains that are out there they’re working on it they’re working on some permissioning the inside of the public blockchain and so are we. Sort of a blockchain to blockchain. You know people talk about side chains in some respects which is a little bit odd. But it will come with blockchain adoption. You know we see customers in kind of three stages of adoption. There are the people who are like hey we’re researching blockchain because my aunt Sue is getting rich off a Bitcoin or because my CIO says you know blockchain is awesome. I need to go research it. So there’s that low tier. There’s kind of a middle tier where they played with it enough to know and understand blockchain and understand some of these things that we’ve been talking about you know the scaling factor or the audit capabilities. Maybe. You know they’re the kind of customer that you have an intelligent conversation with. But then if you say hey if you ever try to do an audit on a transaction and see the true history of a transaction and if they don’t start laughing at you hysterically you know that they really haven’t played with blockchain that much. Because none of the blockchains out there you can really see it true transactional history even though you have this immutable record. And then there’s a third tier where they really come at you and say you know how many nodes do I need. How do I secure these things. How do we get this past my risk management department? How do I operationalize this? There are a very very small number of customers in that top tier very small number of customers and. And those are the customers that really understand the problems that we’re solving in a lot of people we’ll get to that stage. You know like I said it’s an evolutionary process.
But you know I totally forgot the first part of your question but that’s where we see kind of the market going.
[DB] Okay. Now there are many applications and use cases for blockchain technologies. Each time one gets deployed it will have its own chain. However if you were to transfer your information or value from one chain to another that becomes a bit of an issue. Some current solutions for that are to make use of exchanges or to use something like the cosmos network to connect all these various blockchains together. What is the VMware strategy around creating an Internet of blockchains?
[MD] Yeah, we try to take a more intelligent approach at this. There are a lot of projects that are out there and Cosmos is one of them and in the HyperLedger world you’ve got the Quilt folks that are looking at multi- ledgers support.
We truly believe that there’s not going to be one blockchain to rule them all, and that multi-ledger needs to happen. We’re working a lot of the standards bodies to figure out what does this truly mean. Do multi ledger I mean there’s one thing from taking information from one chain and putting into another chain. Pretty much all the solutions that are out there do this through some medium intermediary mechanism.
If it’s a token exchanges or some wallet mechanism if you look at like Microsoft with triplets as an off chain technology wallets are all change technology Cosmos is off chain technology. Basically, it is taking something in and writing it again. Right? How do I know though that that things that changing back and forth actually have wide reaching consensus between two different blockchains that doesn’t exist as has been invented. No one’s really doing that yet. It’s one of the things that we’ve looked at but and had a solution for and stepped back because it’s not something that you can continually guarantee right now until we get to some sort of interlinked protocol that’s standardized. You’re just stitching stuff together and there’s going to be and there’s going to be failures and you can’t protect enough against as failures. You know, blockchain is about this highly distributed high trust environment. Distributed trusts. And you can’t really trust those kind of hacks that are out there. You can get it and you can do a dog and pony show in a demo. But when you really boil it down to and say How can I guarantee the trust of that data and nothing’s been changed. You can’t. And with these intermediary changes so this is something that they’ll be evolving. You know keep an eye out there there are several people looking at it. It’s I think going to be one of the I guess the seventh Grand Challenge in IC3, which is an international academic consortium looking at blockchain technologies that ourselves and many other people in the industry or our members of as well. So you know it’ll get out there but it’s not there yet. And anybody who tells you otherwise has no idea what they’re talking about.
[DB]Okay. Now blockchain on Kubernetes is a VMware Fling that provides some insight for developers building blockchain solutions. Will this eventually end up as part of a platform for enterprises?
[MD]Yes, the Kubernetes work that we did was also because that’s what really started the research from a management perspective of HyperLedger by our China team. So as a couple of our Chinese engineers, Henry Zhang over there, who released this, is kind of the lead engineer on that effort and that’s what got us interested in the management security aspects.
It was a very early prototype of stuff. That deployment engine itself will not make it into our product. We’ve taken an evolutionary step on top of that and we’ve contributed it to the Cello project inside of HyperLedger. So we took that as the basis for our work inside of HyperLedger. So, it is already in there it’s in version I think 1.8 of the Cello release. So it’ll do management deployment of Kubernetes as well as on vSphere, as well as on native inside about that Cello project for HyperLedger. So, use the Fling if you want to you know, its free, its out there it’s unsupportive but it was kind of version 1.0 if you think about it in version 2.0 made it into a much better version made it into the Cello project. And then we are building conversion 3.0 that if you will for our cross blockchain management and security stack that we’re working on right now. So that will manage HyperLedger, and deploy it and do the deployment management of VMwares Enterprise blockchain as well initially.
[DB]So what does a road map of a blockchain related solutions from VMware look like? Are we looking at something that may come out this year? Next year? In a couple of years?
[MD] Yes. You know I can’t give a timeline without risking the revenue for the entire company. But what I can tell you is that we are working hard. Next week is Dell technologies world that when you release this podcast but April 29th is Dell technologies world out in Vegas on the 30th in the morning at eight thirty a.m. If you happen to be in Vegas we are doing a hands on lab with our technology stack. So it’s there and working in functional enough where you can walk through it do a hands on lab.
We will put it up as a public Hands-On lab after that. So this is not you know kind of ghosts, smoke and mirrors and stuff like that. So we’re really far enough along that you can consider “hey we’re putting it out there as a lab”. So use that on your own time scale. Think about when we were we would release stuff from a VMware perspective and you’d probably be close to accurate on building your little time scale without me telling you anything. But it is something that you know we are actively out there with customers. Customers are piloting this technology and using this technology providing feedback. And these are you know the people in that top tier we are talking about. You know the advanced people. Because thats the people were trying to solve the problems for either in scale or in throughput or in time to finality or in security or in multi cloud distributed you know nodes between multiple different clouds. So they’re not locked into a cloud provider and they get a bigger distributed trust. And really what our end game is here is not necessarily blockchain. Blockchains the beginning for this distributed trust. What we’re now doing is concentrating on putting extra services around that distributed trusts that can leverage that distributed trust to build something that we call the trusted data exchange. So stuff like distributed analytics stuff like IoT interfaces things like encrypted search. Or zero knowledge proofs or things, or you know other interesting things that we can do with the distributed trust or center on that trust mantra. And that’s really the trusted data exchange and you’ll hear more about that starting next week at Dell technologies world and continuing on obviously through the summer and through the VMworld timeframe and things like that.
[DB] So, if somebody wanted to follow along with the projects and initiatives that VMware is doing or get access to the hands on labs or keep current on what’s currently going on. What would be the best way for them to do that?
[MD]Luckily the vast majority of stuff that we do in the blockchain group we do through open sharing. Which means that you know, I wear that research hat. So a lot of it’s publish way ahead of before it ever makes it into the product as part research papers that are peer reviewed. You know, we release code thats peer reviewed out there to the academic world.
[MD]So go to research dot VMware dot com. research.vmware.com. Look under the projects tab. You’ll find blockchain right there you’ll find a tonne of whitepapers. They go into excruciating detail on exactly how everything works on our consensus protocol. And and everything else that you might want to find. There is a link there where you can e-mail us and get more information and engage with us that way. We don’t have a public VMware page up just yet. We’re working with MarCom on that but we’re a big company take some time to get that through marketing and get it up there on the public website. So for now go to our research sites you can find us there you find us a lot of different speaking opportunities throughout the summer starting with Dell technologies world. I was in Denver last week at the global blockchain summit out there. So you’ll find more information and then we’ll put the lab link on there when it goes live in the public side of the world as well.
[DB] Wonderful. Well Thank you for taking the time to speak with us today. You’ve provided a lot of really valuable information and I definitely look forward to see what VMware is coming out with in the near future. Thanks a lot.
[MD]Wonderful thank you again for the time. Appreciate it.
References:
https://research.vmware.com/projects/vmware-enterprise-blockchain
Designing Risk in IT Infrastructure 